aix1

45 ms·

601.
by aix1 5y ago
Also this TR: https://www.microsoft.com/en-us/research/wp-content/uploads/...
602.
by aix1 5y ago
In the process of moving from London to Zürich to be closer to mountains and forests of central Europe. (While keeping the same role, now remotely.)
603.
by aix1 5y ago
Can the (edited) headline be clarified to say that the 47% figure is year-in-year ? Otherwise the timelines are not at all clear.
604.
by aix1 5y ago
> by the time there is enough gap for your fingers to slip onto the metal, all the pins will be fully disconnected from the socket I think this should read "by the time the metal parts of the line and neutral pins are exposed, they
605.
by aix1 5y ago
Here's the source code: https://github.com/oklemenz/PrinceJS
606.
by aix1 5y ago
I am not sure that's true anymore. Every single application to work remotely that I know of (including mine[*]) has been approved. That said, most people are choosing to continue being based at whatever office they were based pre-COVI
607.
by aix1 5y ago
And also front-office dev isn't everybody's cup of tea (though I am sure this can be said of any type of role). (Source: worked in finance/trading for 10+ years before jumping ship.)
608.
by aix1 5y ago
I work at one such "top company". Internal mobility is very, very easy and well-supported by the company (great business choice IMO). I am not sure if would even make sense to talk about "sticking around" at the comp
609.
by aix1 5y ago
I'm curious about your definition of huge impact ? (esp. given the timescales.)
610.
by aix1 5y ago
There exist inline insertion signalling standards (e.g. SCTE-35). These could be used for the insertion of ads. However, none of this signalling typically makes it into the final over-the-air or cable broadcast and so is not useful for ad
611.
by aix1 5y ago
Could anyone explain the significance of this finding?
612.
by aix1 5y ago
The time series of frame probabilities is of course highly autocorrelated, but I never bothered to model that due to lack of interest. From memory, a simple single-frame classifier was >95% accurate (this was quite a few years ago, mind
613.
by aix1 5y ago
In terms of eye strain, is that any different to watching PAL TV for five hours straight (25 frames/s or 50 fields/s)?
614.
by aix1 5y ago
A while back I built an ad/not-an-ad classifier for TV frames based purely on pixels (i.e. not prior knowledge of specific ads). One of the interesting findings -- perhaps obvious in retrospect -- was that ad frames were on average br
615.
by aix1 5y ago
I own, or owned, ~10 Anker chargers (various models, bought over the years). At least half them have had at least one port failure after a period of -- very approximately -- 1-2 years. Not a huge deal considering how much use I got out of
616.
by aix1 5y ago
Would this require a local billing address?
617.
by aix1 5y ago
To clarify, both named tuples and dataclasses can be immutable (the former are always immutable and the latter can be made immutable with `frozen=True`). There is no way, however, to make a named tuple mutable .
618.
by aix1 5y ago
What really jumped out at me in your comment is that its definition of fluency (watch tv shows, read books) focuses on comprehension and appears to completely ignore being able to produce anything in the target language.
619.
by aix1 5y ago
Named tuples have been around for a long time (since 2.6), whereas dataclasses are a relatively recent addition to the standard library (3.7). Their differences are highlighted in the dataclasses PEP: https://www.python.org/
620.
by aix1 5y ago
Is there a good "Why Zig" writeup motivating the language? I've been looking around their web site and the only thing I could find was pretty generic: Zig is a general-purpose programming language and toolchain for maintaini
621.
by aix1 5y ago
This is a neat trick, no question. It is only about the performance of the decorator itself, not about the decorated class, right? So faster imports (if imports are your bottleneck) but equally-performant classes.
622.
by aix1 5y ago
Thanks for sharing. Here is another paper on the subject (cited in the Johns Hopkins study you linked to): https://journals.lww.com/journalpatientsafety/Fulltext/2013/... > Using a weighted average of th
623.
by aix1 5y ago
Firstly, "3rd most common cause of death" != "a third of all people". But putting that aside, I too found the claim implausible and went looking for data. Here is what the CDC has to say: Leading Causes of Death D
624.
by aix1 5y ago
Obligatory: What Every Computer Scientist Should Know About Floating-Point Arithmetic https://docs.oracle.com/cd/E19957-01/806-3568/ncg_goldberg.h...
625.
by aix1 5y ago
I don't know if my experience is representative, but I'm well over 40 and haven't felt this one bit.
626.
by aix1 5y ago
A couple of points: 1. It's easy to forget how recent many of modern ML methods for computer vision are. (E.g. U-net only goes back to 2015!) 2. It's not totally clear to be what you mean by "mixed results" (have we sol
627.
by aix1 5y ago
Your comment reminded me of this fascinating work: Levenson RM, Krupinski EA, Navarro VM, Wasserman EA (2015) Pigeons (Columba livia) as Trainable Observers of Pathology and Radiology Breast Cancer Images. PLoS ONE 10(11): e0141357. http
628.
by aix1 5y ago
You're exactly right: by construction, IQ scores are normally distributed with a mean of 100 and a standard deviation of 15.
629.
by aix1 5y ago
So "least affordable" (median house price as a multiple of median income). To me that's quite different from "most expensive" (highest dollar prices).
630.
by aix1 5y ago
Berlin, Dublin, Zürich. Pay & benefits are adjusted to be in line with anyone else employed by the company at the target office. Being able to pick a city where one actually wants to live is definitely an upside. Being a remote sing

More ›